00spatialkernel-package: The Spatialkernel Package

In spatialkernel: Nonparametric Estimation of Spatial Segregation in a Multivariate Point Process

Description

An R package for spatial point process analysis.

Details

This package contains functions for spatial point process analysis using kernel smoothing methods. This package has been written to be compatible with the splancs package which is available on CRAN (The Comprehensive R Archive Network).

For a complete list of functions with individual help pages, use library(help = \ "spatialkernel").

Maintainer

Pingping Zheng [e-mail]

Note

For the convience of the user, we present here examples which show how to use some of the functions in the package.

Author(s)

Pingping Zheng and Peter Diggle

References

1. P. Zheng, P.A. Durr and P.J. Diggle (2004) Edge-correction for Spatial Kernel Smoothing — When Is It Necessary? Proceedings of the GisVet Conference 2004, University of Guelph, Ontario, Canada, June 2004.

2. Diggle, P.J., Zheng, P. and Durr, P. A. (2005) Nonparametric estimation of spatial segregation in a multivariate point process: bovine tuberculosis in Cornwall, UK. J. R. Stat. Soc. C, 54, 3, 645–658.

See Also

cvloglk, phat, mcseg.test, plotphat, plotmc, pinpoly, risk.colors, metre

Examples

## An example of spatial segregation analysis
## Not run: 
## source in Lansing Woods tree data within a polygon boundary
data(lansing)
data(polyb)
## select data points within polygon
ndx <- which(pinpoly(polyb, as.matrix(lansing[c("x", "y")])) > 0)
pts <- as.matrix(lansing[c("x", "y")])[ndx,]
marks <- lansing[["marks"]][ndx]
## select bandwidth
h <- seq(0.02, 0.1, length=101)
cv <- cvloglk(pts, marks, h=h)$cv
hcv <- h[which.max(cv)]
plot(h, cv, type="l")
## estimate type-specific probabilities and do segregation tests
## by one integrated function
sp <- spseg(pts, marks, hcv, opt=3, ntest=1000, poly=polyb)
## plot estimated type-specific probability surfaces
plotphat(sp)
## additional with pointwise significance contour lines
plotmc(sp, quan=c(0.025, 0.975))
## p-value of the Monte Carlo segregation test
cat("\np-value of the Monte Carlo segregation test", sp$pvalue)

##estimate intensity function at grid point for presentation
##with bandwidth hcv
gridxy <- as.matrix(expand.grid(x=seq(0, 1, length=101), y=seq(0, 1, length=101)))
ndx <- which(pinpoly(polyb, gridxy) > 0) ##inside point index
lam <- matrix(NA, ncol=101, nrow=101)
lam[ndx] <- lambdahat(pts, hcv, gpts = gridxy[ndx,], poly =
    polyb)$lambda
brks <- pretty(range(lam, na.rm=TRUE), n=12)
plot(0, 0, xlim=0:1, ylim=0:1, xlab="x", ylab="y", type="n")
image(x=seq(0, 1, length=101), y=seq(0, 1, length=101),
    z=lam, add=TRUE, breaks=brks, col=risk.colors(length(brks)-1))
polygon(polyb)
metre(0, 0.01, 0.05, 0.51, lab=brks, col=risk.colors(length(brks)-1), cex=1)

## An example of inhomogeneous intensity function and K function
## estimated with the same data
s <- seq(0, 0.06, length=101)
lam <- lambdahat(pts, hcv, poly=polyb)$lambda
kin <- kinhat(pts, lam, polyb, s)
plot(kin$s, kin$k-pi*(kin$s)^2, xlab="s", ylab="k-pi*s^2", type="l")

## End(Not run)

spatialkernel documentation built on May 2, 2019, 4:37 p.m.